Path tracking control for car-like robots by pid controller with time-varying parameters
Khanh Ly Thi Trinh,Hong Thai Nguyen,,
DOI: https://doi.org/10.32508/stdjet.v5i3.961
2022-01-01
Science & Technology Development Journal - Engineering and Technology
Abstract:Wheel mobile robots have been widely applied to many areas of life, from serving people in services and industrial logistics to military fields. Applications of wheeled mobile robots such as floor cleaning robots, supermarket robots, logistics robots in hospitals, autonomous robots that guide automatic guidance in intelligent factories, military robots and agricultural robots. Thereby, we can see the great potential of mobile robots in the future. However, among the wheel mobile robots, the differential-drive mobile robot structure is most commonly used because of its simplicity in structure and control. Nevertheless, the navigation and driving are done synchronously through the two drive wheels, which leads to slippage and difficulty in controlling posture errors when navigating, especially on roads curved with a small radius. In order to overcome the disadvantages of the steering and navigation of differential-drive mobile robots, the structure of the automobile has been applied to design a wheel-mobile robot called a car-like robot. Many studies have shown that the advantage of this design option is that the navigation control and robot drive are separated. As a result, the dynamic control loop can ignore the navigation part, while the kinematic model determines the navigation law. Therefore, the work proposes the design of a PID controller with time-varying parameters for a car-like robot to track a given trajectory with minimum error. The nonlinear kinematic model of the robot is linearized along a reference trajectory, and the obtained linear model is used in the controller design process. A PID controller is designed where the controller parameters are tuned to minimize the tracking error. Our model achieves a tracking error value, including the minor position is 5.1 mm, and the maximum postural error is 2o. The simulation results showed the proposed controller's effectiveness in position and posture errors, and it shows we can apply this research to control car-like mobile robots in logistics services.